HIGH-PERFORMANCE CONCRETE FOR HIGHWAY STRUCTURES A GENERIC REVIEW OF DEFINITIONS, MIXTURE PROPORTIONS AND PERFORMANCE LEVELS

Considered are commonly used constituent materials and their quantities, cementitious material combinations, and performance levels typically achieved.. Particular HPC characteristics of

HIGH-PERFORMANCE CONCRETE FOR HIGHWAY STRUCTURES:

A GENERIC REVIEW OF DEFINITIONS, MIXTURE PROPORTIONS AND

PERFORMANCE LEVELS

Bryan J Magee Research Assistant Professor, University of New Hampshire

Durham, New Hampshire, U.S.A

Jan Olek Associate Professor, Purdue University West Lafayette, Indiana, U.S.A

ABSTRACT

This paper summarizes an extensive review of literature that was undertaken to better quantify the term High-Performance Concrete (HPC) Based on a database of 260 HPC mixtures, descriptions of typical HPC characteristics are given Considered are

commonly used constituent materials and their quantities, cementitious material

combinations, and performance levels typically achieved It is illustrated that a

quantitative HPC definition as proposed by the FHWA is generally representative of HPC used in the field to date, providing therefore, a rational definition of the material Against this background, the data reported in this paper provides an approximate trial mixture-proportioning guide for HPC

INTRODUCTION

At the outset of an ongoing research project at Purdue University, the main goal of which

is to develop procedures that will yield concrete bridge superstructures of improved and predictable performance, High-Performance Concrete (HPC) was identified as one of several construction materials meriting further investigation Indeed, based on research undertaken throughout the past decade, HPC has already been adopted for numerous bridge applications in recent times owing to both its improved strength development and durability characteristics1-3 Despite this trend, however, HPC is not a well-defined material, and for many exists as somewhat of an enigma In fact, it has been recently stated 4 that while to many engineers the term HPC is recognizable, providing a definition for the material on paper presents considerable difficulty For this reason, the question asked at the inception of the current research project was simply; what is HPC? Against this background, the work reported in this paper was carried out in an attempt to

demystify the term HPC by reviewing existing definitions and examining typical material characteristics Particular HPC characteristics of interest included parameters such as: (i)commonly used constituent materials and their quantities; (ii) cementitious material combinations; and (iii) performance levels typically achieved To examine these issues,

an extensive review of pertinent literature has been carried out, concentrating solely on construction or research projects reporting the use of HPC

REVIEW OF EXISTING HPC DEFINITIONS

The lack of a universally accepted definition for HPC is, without doubt, a significant contributor to the uncertainty surrounding the material From the literature, it is clear that two distinct classifications of HPC definitions have emerged over the years; namely descriptive and quantitative-style definitions

Descriptive HPC Definitions

Owing to the inherent vagueness of the term HPC, it is often opinioned that the material can only be defined by referring to the performance requirements of the intended use of the concrete As a result, many HPC definitions are unspecific and are proffered without making any reference to specific concrete characteristics For example, the American Concrete Institute 4 defines HPC as “…concrete meeting special combinations of

performance and uniformity requirement that cannot always be achieved routinely using conventional constituents and normal mixing, placing, and curing practices” Similarly, Swamy 5 defines HPC as that “…designed to give optimized performance characteristics for a given set of loads, usage, and exposure conditions consistent with the requirements of cost, service life, and durability” At the same time, however, many descriptive HPC definitions do refer to specific concrete characteristics For example, Neville 6 states that

“…the essential feature of HPC is that its ingredients and proportions are specifically chosen so as to have particular appropriate properties for the expected use of the structure;

these properties are usually a high strength or low permeability” Similarly, the FHWA

have reported 7 that “HPC is concrete that has been designed to be more durable and, if necessary, stronger than conventional concrete” A more comprehensive review of

descriptive HPC definitions may be found in a recent paper in Concrete International 4

Quantitative HPC Definitions

In contrast, attempts have been made to define HPC quantitatively For example, Aitcin 8 defined HPC as essentially all concrete with a low water-binder ratio ( 0.40) The Strategic Highway Research Program (SHRP) defined HPC as having (i) a maximum water-binder ratio of 0.35, (ii) a minimum durability factor of 80% (as determined by ASTM C 666 Method A), (iii) a minimum compressive strength of either 21 MPa (3046 psi) within 4 hours, or 34 MPa (4931 psi) within 24 hours, or 69 MPa (10,008 psi) within

28 days 9 More comprehensively, the Federal Highway Administration (FHWA) has proposed a HPC definition based on performance criteria that includes four HPC grades, eight performance characteristics, and four exposure conditions10 As shown in Table 1, the four HPC grades are related to the severity of exposure, and recommended for each grade are both performance characteristic values and procedures to measure performance

DETAILS OF DATA COLLECTION AND ANALYSIS

Although numerous individual information sources related to HPC were located in the literature, few of these met with the objective of this paper, which is to provide a generic overview of the material from a national perspective In this way, the intention is to

TABLE 1 – Proposed FHWA definition of HPC for structural concrete 10

Exposure Conditions /

Performance Characteristics

Standard Test Method

FHWA HPC Performance Grade Grade 1 Grade 2 Grade 3 Grade 4

Exposure Conditions

 Freeze-thaw durability exposure

- Scaling resistance, applied salt

- Abrasion resistance

(X = average daily traffic) - X  50,000 50,000 < X < 100,000 100,000  X

- Chloride resistance, applied salt

(X = tons / lane-mile-year) - 1.0  X < 3.0 3.0  X < 6.0 6.0  X

-Performance Characteristics

 Freeze-thaw durability (300 cycles)

(X = dynamic modulus, %)

AASHTO T161

- Scaling resistance (50 cycles)

(X = visual surface rating) ASTM C 672 X = 4.5 X = 2.3 X = 0.1

- Abrasion resistance

(X = average wear depth, mm) ASTM C 944 2.0 > X  1.0 1.0 > X  0.5 0.5 > X

- Chloride resistance

(X = Columbus)

AASHTO T 277 ASTM C 1202

3000  X >

2000 2000  X > 800 800  X

- Compressive strength

(X = 28 day result, MPa)

AASHTO T 2 ASTM C 39 41  X < 55 55  X < 69 69  X < 97 X  97

 Modulus of elasticity

(X = 28 day result, GPa) ASTM C 469 28  X < 40 40  X < 50 X  50

- Shrinkage

(X = Microstrain) ASTM C 157 800 > X  600 600 > X  400 400 > X

- Creep

(X = Microstrain / MPa) ASTM C 512 75  X > 60 60  X > 45 45  X > 30 30  X

provide guidance to practicing concrete engineers and technicians attempting to design HPC mixtures To meet these objectives, an extensive literature review was performed using resources of Purdue University library; SHRP concrete and structures program resources (materials from HPC showcases); TRIS literature database; as well as Office of Technology Application, PCI and ACI on-line search engines In addition, information was also provided from State materials engineers The sole criterion used when identifying relevant literature and information sources was reference to the term ‘HPC’ in the article title or abstract This was based on the premise that all authors applied some qualitative or quantitative criterion when classifying their concrete as being ‘high-performance’ In total, around 200 recent literature sources describing either research-related topics or practical applications associated with HPC were located By extracting relevant information

available from each, a database of 260 HPC mixtures was subsequently compiled Where available for each mixture, the information recorded included: (i) constituent materials and their quantities; (ii)cementitious material combinations used; and (iii)performance levels achieved The important characteristics of the numerical data collected were then analyzed and summarized using traditional descriptive statistical methods (such as the measuring the data center, spread and variation) The trends obtained from this work are discussed in the following sections

TYPICAL HPC CHARACTERISTICS Constituent Material Quantities

In order to identify typical HPC compositions, a compilation of histograms was developed

to highlight the constituent material quantities most frequently used for HPC (see

Figure1) Shown in Figures1(a)-(e) respectively are the most commonly used water, total binder, air, fine and coarse aggregate contents for HPC Each of the figures also includes statistical information such as mean, standard deviation and coefficients of

variation

Clearly the use of relatively low water contents (150-175 kg/m3) (252-295 lb/yd3) and high total binder contents (350-500 kg/m3) (589-841 lb/yd3) are most common for HPC It is interesting to note that while the most frequently used total binder content range was

350-400 kg/m3 (589-673 lb/yd3), a second peak range of 450-500kg/m3 (757-841 lb/yd3) was also identified in Figure 1(b) The occurrence of two peak ranges is likely a reflection of mixtures with and without supplementary cementitious materials Interestingly, the

minimum and maximum values reported for water content and total binder content

respectively were 56 and 675 kg/m3 (95 and 1136 lb/yd3) although not surprisingly, these values were not reported for the same mixture! Similar to the trend noted for total binder content, two air content peak ranges of 1-2% and 5-6% were identified for HPC (see Figure1(c)) most likely reflecting mixtures with and without the use of air entraining chemical admixtures respectively Not dissimilar to quantities used for conventional concrete, the most commonly used fine and coarse aggregate ranges for HPC were 700-800 and 1000-1100 kg/m3 (1178-1345 and 1683-1851 lb/yd3) respectively Information

regarding chemical admixture types and quantities typically used for HPC is not included

in Figure1 for clarity, and owing to the fact that dosages and units in which dosages were reported varied widely in the literature In summary, however, admixtures most frequently used for HPC as reported in the literature included high and normal range water-reducers, air entrainers, and retarders High range water-reducers (superplasticizers) were by far the most frequently used group of admixtures, being specified for more than 50% of the mixes reviewed Dosages of superplasticizer were typically high, ranging from around

5-15liters/m3 (3.8-11.5 litres/yd3) of concrete

Binder Material Combinations

In addition to Portland cement, three supplementary cementitious materials have

predominantly been used for HPC; namely silica fume (SF), fly ash (FA) - Class C and F, and ground granulated bastfurnace slag(GGBS) As shown in Figure 2, six different binder combinations have been used based around these four binder materials Used for 46% of all mixtures reviewed, PC/SF combinations are by far the most common for HPC Silica fume has also been used in ternary blends with fly ash (PC/FA/SF) and granulated-blastfurnace slag (PC/GGBS/SF), with these combinations accounting for 17 and 7% of mixtures respectively The remaining 15, 8 and 7% of binder combinations used for HPC were PC alone, and the binary blends PC/FA and PC/GGBS respectively

0 10 20 30

Air content, %

0

40

80

0 20 40 60 80

0

40

80

120

160

0 20 40 60 80

Coarse agg.

Fine agg.

Air Binder

Water

(b)

(c)

(d)

Mean: 443 Total range: 222-675 S.D.: 79.0

C of V.: 17.9%

Mean: 144

Total range: 56-221

S.D: 20.5

C of V.: 14.2

Mean: 4.0 Total range: 0.8-9.8 S.D.: 2.6

C of V.: 64.6%

Mean: 1081 Total range: 561-1608 S.D.: 168.0

C of V.: 15.5%

Mean: 720 Total range: 318-1205 S.D.: 139.4

C of V.: 19.4%

Binder material quantities

To further clarify the typical binder composition of HPC, a compilation of histograms was developed (see Figure 3) to summarize commonly used water-binder ratios and binder material quantities for HPC Binder material quantities in this figure are expressed as a percentage by mass of the total binder content used for each mixture

The use of relatively low water and high total binder contents for HPC (as already shown

in Figure 1) is clearly reflected in Figure 3(a), which indicates 0.25 to 0.40 as being the most commonly used water-binder ratio range for HPC The lowest reported water-binder ratio value was 0.19, while only very infrequently were values greater than around 0.50 used

The use of various PC replacement materials in HPC (see Figure 2) is clearly reflected in Figure 3(b), which indicates that PC contents as low as 22% by mass of total binder have been used Clearly though, the most commonly used PC content range for HPC is 90-100% by mass of total binder This trend reflects the fact that PC only and PC/SF binder combinations were reported for 15 and 46% of HPC mixtures respectively, with the latter tending to use replacement levels in the range 0-20% by mass (see Figure1(c)) It should

be noted that while a SF replacement as high as 60% by mass of total binder content was used, values higher than 20% were used very infrequently As is typical for conventional concrete, FA replacement levels ranged from 0 to 45% by mass of total binder content, see Figure 3(d) In this range, no particular replacement level stood out as being most popular

In comparison, two peak ranges of GGBS are identifiable in Figure3(e) While the peak GGBS range reported was 0-5% by mass of total binder content, the majority of

replacement levels fell in the range 25-50%, with 35-40% being the second most

commonly used range A review of the database showed that the 0-5% range represented mixtures using small quantities of GGBS in ternary combinations with PC and SF

46 %

17 % 15 %

0 10 20 30 40 50

PC/S F

PC/S

PC/F

A

PC/G BS

PC/S

GBS

Binder material combination

GGBS: Ground granulated blastfurnace slag

Figure 2 Binder material combinations most frequently used for HPC

Figure 3 HPC binder material percentages most frequently used (information taken from database of 260 mixtures)

Air

BINDER

Water 0

40

80

120

160

20 -

30

30 -

40

40 -

50

50 -

60

60 -

70

70 -

80

80 -

90

90 -

100

PC content, % by mass of total binder content

0

40

80

120

160

0 - 5

10 -

20

20 -

25

30 -

35

40 -

45

50 - 55

SF content, % by mass of total binder content

0 2 4 6 8 10

0 - 5

10 -

15

20 -

25

30 -

35

40 -

45

50 -

55

60 - 65 GGBS content, % by mass of total bind er content

0 4 8 12 16 20 24

0 - 5 5 - 10

10 -

15

15 -

20

20 -

25

25 -

30

30 -

35

35 -

40

40 - 45

FA content, % by mass of total bind er content

(b)

(c)

(d)

(e)

Mean: 84.5 Total range: 22-100 S.D.: 18.7

C of V.: 254 %

Mean: 6.7 Total range: 0-60 S.D.: 7.6

C of V.: 113.7%

Mean: 4.6 Total range: 0-45 S.D.: 9.8

C of V.: 212%

Mean: 4.2 Total range: 0-65.0 S.D.: 13.0

C of V.: 310%

0 20 40 60 80

0.15

0.20 0.20

0.25 0.25

0.30 0.30

0.35 0.35

0.40 0.40

0.45 0.45

0.50 0.50

0.55 0.55

0.60 0.60

0.65 0.65

0.70

W ater / binder ratio

Total range: 0.19-0.70 S.D.: 0.10

C of V.: 21.9%

FA

GGBS SF

PC

HPC Constituents +

Figure 3 HPC binder material percentages most frequently used (information taken from database of 260 mixtures)

Air

BINDER

Water 0

40

80

120

160

20 -

30

30 -

40

40 -

50

50 -

60

60 -

70

70 -

80

80 -

90

90 -

100

PC content, % by mass of total binder content

0

40

80

120

160

0 - 5

10 -

20

20 -

25

30 -

35

40 -

45

50 - 55

SF content, % by mass of total binder content

0 2 4 6 8 10

0 - 5

10 -

15

20 -

25

30 -

35

40 -

45

50 -

55

60 - 65 GGBS content, % by mass of total bind er content

0 4 8 12 16 20 24

0 - 5 5 - 10

10 -

15

15 -

20

20 -

25

25 -

30

30 -

35

35 -

40

40 - 45

FA content, % by mass of total bind er content

(b)

(c)

(d)

(e)

Mean: 84.5 Total range: 22-100 S.D.: 18.7

C of V.: 254 %

Mean: 6.7 Total range: 0-60 S.D.: 7.6

C of V.: 113.7%

Mean: 4.6 Total range: 0-45 S.D.: 9.8

C of V.: 212%

Mean: 4.2 Total range: 0-65.0 S.D.: 13.0

C of V.: 310%

0 20 40 60 80

0.15

0.20 0.20

0.25 0.25

0.30 0.30

0.35 0.35

0.40 0.40

0.45 0.45

0.50 0.50

0.55 0.55

0.60 0.60

0.65 0.65

0.70

W ater / binder ratio

Total range: 0.19-0.70 S.D.: 0.10

C of V.: 21.9%

FA

GGBS SF

PC

HPC Constituents +

Air

BINDER

Water 0

40

80

120

160

20 -

30

30 -

40

40 -

50

50 -

60

60 -

70

70 -

80

80 -

90

90 -

100

PC content, % by mass of total binder content

0

40

80

120

160

0 - 5

10 -

20

20 -

25

30 -

35

40 -

45

50 - 55

SF content, % by mass of total binder content

0 2 4 6 8 10

0 - 5

10 -

15

20 -

25

30 -

35

40 -

45

50 -

55

60 - 65 GGBS content, % by mass of total bind er content

0 4 8 12 16 20 24

0 - 5 5 - 10

10 -

15

15 -

20

20 -

25

25 -

30

30 -

35

35 -

40

40 - 45

FA content, % by mass of total bind er content

(b)

(c)

(d)

(e)

Mean: 84.5 Total range: 22-100 S.D.: 18.7

C of V.: 254 %

Mean: 6.7 Total range: 0-60 S.D.: 7.6

C of V.: 113.7%

Mean: 4.6 Total range: 0-45 S.D.: 9.8

C of V.: 212%

Mean: 4.2 Total range: 0-65.0 S.D.: 13.0

C of V.: 310%

0 20 40 60 80

0.15

0.20 0.20

0.25 0.25

0.30 0.30

0.35 0.35

0.40 0.40

0.45 0.45

0.50 0.50

0.55 0.55

0.60 0.60

0.65 0.65

0.70

W ater / binder ratio

Total range: 0.19-0.70 S.D.: 0.10

C of V.: 21.9%

FA

GGBS SF

PC

HPC Constituents +

Performance levels

Although numerous publications dealing with HPC can be found, these typically only provide data on a limited number of properties Properties reported frequently enough to allow representative statistical analysis included: slump; compressive strength; modulus of elasticity; and chloride ion permeability Perhaps not surprisingly, 28-day compressive strength was the most frequently reported property, with 89% of the reviewed literature sources reporting a result In comparison, slump, chloride ion permeability and modulus

of elasticity were reported in 51, 33 and 10% of cases respectively Given in Figure 4 is a further suite of histograms, in this case highlighting the most commonly reported HPC performance levels

In terms of slump, values reported for HPC were generally high, with 150-200 mm (5.9-7.9in.) being the most common range (Figure4(a)) Despite a slump as low as 6-mm being reported, the generally high slump values reported were attributed to the use of high dosages of plasticizing chemical admixture In terms of mechanical properties, the most common modulus of elasticity and 28-day compressive strength ranges reported for HPC were 35-40 GPa and 50-100MPa (5,075-5,800 ksi and 7,251-14,504 psi), respectively (Figure 4(b) and (c)) Indeed, with 28-day strength values as high as 143 MPa (20,735psi) reported, these figures go some way to substantiating the frequent preconception that

‘high-performance’ and ‘high-strength’ concretes are cognate This was not always the case, however, with selective mixtures with 28-day compressive strength values as low as

30 MPa (4,350psi) being reported as being HPC In many cases, therefore, it is clearly not the development of high levels of compressive strength that defines ‘high-performance’ Instead, and as illustrated in Figure 4(d), the criterion defining ‘high-performance’ in many instances is clearly the attainment of low levels of chloride ion permeability Measured following the procedures of AASHTO T-277, a high majority of the HPC mixes reported

in the literature achieved 28-day chloride ion penetration values lower than 2000

Coulombs, with 500-1500 being the most commonly reported range Using the criteria provided in ASTMC1202-94, all of these concretes may be classified as having low to negligible chloride ion permeability

Disparity between ‘high-performance’ and ‘high-strength’ concrete is further illustrated in Figure 5, which shows the relationship between 28-day compressive strength and chloride ion permeability for those HPC mixtures where both values were reported in the literature Despite showing a general trend of decreasing chloride ion permeability with increasing compressive strength (r2 = 0.54), it is evident from Figure 5 that attaining

‘high-performance’ in terms of permeability (i.e <1000 Coulombs) is not exclusive to mixtures attaining ‘high-strength’ Indeed, Coulomb values lower than 1000 have been reported for mixtures with 28-day compressive strength values ranging from 32-89MPa

(4,640-12,905psi) In this way, Figure 5 substantiates the legitimate opinion that concrete strength development should not be viewed as the only suitable indicator of concrete durability

20

40

60

Slump, mm

Figure 4 Typical HPC properties (information taken from database of 260 mixtures)

[1mm ~ 0.039in.; 1GPa ~ 145 ksi; 1MPa ~ 145 psi]

(a)

Mean: 147

Total range: 6.0-270

S.D.: 72.0

C of V.: 48.9%

0 20 40 60 80 100

28-day compressive strength, MPa

(c)

Mean: 74 Total range: 25-143 S.D.: 25.6

C of V.: 34.8%

0 2 4 6 8 10

28-day modulus of elasticity, GPa

(b)

Mean: 38.0 Total range: 24.0-60.0 S.D.: 6.5

C of V.: 16.8%

0 5 10 15 20 25

28-day chloride ion permeability, Coulombs

(d)

Mean: 1522 Total range: 115-7460 S.D.: 1349.9

C of V.: 88.7%

Selected HPC Properties

Durability Fresh

Mechanical

SUMMARY AND CONCLUSIONS

Practical Significance of Descriptive and Quantitative HPC Definitions

It has been shown that HPC definitions are either descriptive or quantitative in nature Descriptive definitions have been developed owing to the common belief that there is no unique definition of HPC and that as stated by Rangan 11, “…HPC can be defined only with reference to the performance requirements of the intended use of the concrete” In this way, the advantageous feature of descriptive HPC definitions is that they are

universally acceptable, being relevant for all types of concreting applications In

comparison, however, quantitative HPC definitions are typically developed with a

particular concrete application in mind For instance, the quantitative definition developed

by Goodspeed et al 10 has been adopted by the FHWA for concrete used in bridges The numerical values of strength and durability parameters used in this definition have already been presented in Table 1 While not having universal relevance, a major advantage of quantitative HPC definitions is clearly that they are suitable for incorporation in

performance-based or performance-related specifications (PRS) Indeed, an ability to provide numerical values for performance characteristics is the premise for successful implementation of PRS With this said, the co-existence of descriptive and quantitative HPC definitions seems justified, provided the limitations and merits of each are realized While tailored for HPC use in bridge applications, the HPC grades proposed by the FHWA are fairly representative of performance levels that have been achieved in practice for a wide range of concreting applications This is illustrated in Figure 6, where selected

FHWA grades (as previously described in Table 1) have been superimposed onto the HPC performance levels most commonly reported in the literature (as previously illustrated in Figure 4) In this way, it can be seen that for the concrete properties considered

(i.e.modulus of elasticity, compressive strength, and chloride ion permeability), actual performance levels as measured in practice coincide closely with the mid-range of

performance grades as proposed by the FHWA

0 2000 4000 6000 8000

28-day strength, MPa

Area corresponding to ‘low’ and

‘very low’ chloride ion permeability (ASTM C 1202-94)